cognix 0.2.1

AI-powered CLI development assistant

Cognix

Autonomous code generation powered by flow engineering.

---

Quick Start

1. Install and run

pipx install cognix
cognix

2. First-time setup

When you run Cognix for the first time, an interactive wizard will help you set up your API key:

• Choose your AI provider (Anthropic, OpenAI, or OpenRouter)
• Enter your API key
• The wizard creates a .env file automatically

3. Generate code

Try the included sample first (use @ to specify a file):

cognix> /make @sample_spec_tetris.md

Or describe what you want to build:

cognix> /make "landing page with HTML and CSS"

A sample specification file sample_spec_tetris.md is included in the repository. Use it as a reference for writing your own specifications.

4. Available commands

Type /help in the CLI to see all available commands.

---

API Key Setup

Automatic setup (recommended)

Just run cognix and follow the interactive wizard.

Manual setup

Create a .env file in your project directory:

Anthropic Claude (default):

ANTHROPIC_API_KEY=sk-ant-your_key_here

Get your key at: https://console.anthropic.com/

Supported models: Sonnet 4.5 (default), Opus 4.5

OpenAI:

OPENAI_API_KEY=sk-your_key_here

Get your key at: https://platform.openai.com/api-keys

Supported models: GPT-5.2, GPT-5.2 Codex

OpenRouter:

OPENAI_API_KEY=sk-or-v1-your_key_here
OPENAI_BASE_URL=https://openrouter.ai/api/v1

Get your key at: https://openrouter.ai/keys

Switch models

cognix> /model

---

MCP Server Integration

Use Cognix from Claude Desktop, Cursor, VSCode, or any MCP-compatible tool.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "cognix": {
      "command": "cognix-mcp"
    }
  }
}

---

Data Storage

Cognix stores data in ~/.cognix/:

~/.cognix/
├── config.json            # Your settings
├── memory.json            # Conversation & project memory
├── repository_data.json   # Repository analysis cache
├── ui-knowledge.json      # UI component knowledge
├── sessions/              # Saved work sessions
├── knowledge/             # App pattern definitions
├── rules/                 # File reference rules
├── backups/               # Automatic backups
└── impact_analysis/       # Code impact analysis results

Privacy: No telemetry. API calls only go to your configured LLM provider.

---

System Requirements

• OS: Windows 10+, macOS 10.15+, or Linux
• Python: 3.9 or higher
• Internet: Required for LLM API access

---

Links

• Documentation: github.com/cognix-dev/cognix
• Report Issues: GitHub Issues
• Discussions: GitHub Discussions

---

License

Apache-2.0 License - see LICENSE file for details